LIBRARY STATE P£\4$T9B£>ARD E _ 841 United States Department of Agriculture Agricultural Research Administration Bureau of Entomology and Plant Quarantine THE LARCH SAWFLY EPIDEMIC IN THE LAKE STATES Surveys and Spraying Experiments in 1949 and 1950^ 9/ By James W. Butcher-^ and Charles B. Eaton, Division of Forest Insect Investigations The larch sawfly (P ristiphora erichsonii (Htg.)), a well-known enemy of tamarack (Larix laricina (Du Roi) K. Koch), has once again appeared in epidemic numbers in the Lake States. Since the destructive outbreaks of the 19th and early part of the 20th century, during which the insect is reputed to have spread westward from the eastern seaboard to British Columbia and the Yukon (Swaine 5), this sawfly has remained innocuous in the young stands that have developed in the wake of these outbreaks. Its resurgence in the present epidemic apparently began in northern Minnesota as early as 194C (Hodson 2). The pattern of build-up in the extensive tamarack stands in the State between 1940 and 1947 has been charted by Hodson and Christensen (3). In 1949 and 1950 investigations of the larch sawfly were undertaken by the Milwaukee, Wis., laboratory of the Bureau of Entomology and Plant Quarantine. This action was prompted by the inadequacy*- of existing information on the status of the outbreak, the biology of the insect, and the potentialities for destruction in the stands under attack. The objectives were (1) to conduct periodic surveys of the infestation, (2) to study the biology and ecology of the insect, and (3) to evaluate its susceptibility to control with recently developed insecticides. This paper reports tech- niques and results of the surveys and control experiments carried out during these years. 1/ This work was conducted in cooperation with the Division of Ento- mology and Economic Zoology, University of Minnesota. Facilities and assistance were supplied by the U. S. Forest Service; the Minnesota Department of Conservation and Department of Agriculture, Dairy, and Food; and the Office of the St. Louis County (Minnesota) Land Commis- sioner. 2/ Now with Office of the State Entomologist, Minnesota Department of Agriculture, Dairy, and Food. -2- Aerial Reconnaissance The survey work carried out in 1949 was necessarily less thorough than was desirable. Limitations in personnel and time, together with the size of the area to be covered (about 15 million acres in Minnesota alone), made it impossible to survey infested stands adequately from the ground. ' > Fortunately, however, exploratory airplane flights disclosed the feasibility of delimiting defoliated stands from the air and, despite its obvious short- comings, this technique became the chief means of defoliation detection. ' In 1949 an observer and a mapper accompanied the pilot on most of the survey flights. At an air speed of 70 miles per hour and an altitude of 1, 000 feet, it was possible to detect the more obviously defoliated areas in a strip up to 8 miles wide (4 miles on each side of the plane), and to plot them on a map at a scale of 1 inch = 1 mile. It is calcu- lated that the area observed during the 12 hours' flying time devoted to this work was about 4,300,800 acres. The information obtained was checked against aerial photographs of forest type supplied by the United States Forest Service and the Office of the St. Louis Countv (Minnesota) Land Commissioner. Because of the concentric pattern of defoliation, particularly in the larger bogs with extremely wet interiors, it was not possible to convert the survey data into graduated infestation levels. However, the method did permit the spotting of centers of heavy defoliation and provided a basis for the construction of a map illustrating the apparent limits of the 1949 infestation. Visible defoliation was clas- sified as follows: Light, less than 25 percent; medium, 26 through 50 percent; and heavy, 51 through 100 percent. The conduct of aerial reconnaissance surveys in 1950 differed little from the method used in 1949. However, since there was more time for ground surveys, it was possible to forego detailed mapping from the air. Because the chief object was to confirm the pattern of the outbreak as constructed from plot records, flights were carried out at an altitude of 2,000 feet to increase the area of coverage. In this manner at least 6,567,680 acres were examined for evidence of defoliated tamarack. As there was no appreciable extension in 1950 of the area of heavy 1949 defoliation, it is possible to present the results of the 1949 and 1950 surveys in one map (fig. 1). The only apparent difference in con- ditions for the 2 years occurred in the north-central section of Min- nesota (zone 4), where a marked decline in defoliation was observed in 1950. This section embraces most of the tamarack within the Chippewa National Forest. The recession is attributable both to excessive rainfall during the period of cocooning in the summer of 1949, when large num- bers of mature and newly cocooned larvae were drowned, and to the -3 L~ o 05 T— I I c <u -t-» a c o •rH CO +-> W OJ (m C cd w o u a 0) u •iH fa - 4- prolonged immersion of overwintered prepupae^r as a consequence of extremely high water levels in most of the larger bogs in this area during the spring of 1950. One or both of these conditions drastically reduced the population of sawfly larvae on the Chippewa National Forest and its environs during this period. Ground Reconnaissance In 1950 a modified roadside-sampling technique was devised, and the location of each observation point was recorded as the speedometer dis- tance from a prominent check point along well-marked roads. A total of 36 plots were established, of which 23 were located in Minnesota, chiefly on the Superior and Chippewa National Forests, 4 in northern Wisconsin, and 9 in Michigan (fig. 1). Data on tree mortality and degree of defolia- tion, based on examination of at least 10 randomly selected trees per plot, were obtained together with routine observations on stand and site characteristics. These plots and records will form the basis for sub- sequent surveys as long as the outbreak coutinues. In 1949 and 1950 the outbreak was confined to 21 northern Minnesota counties, but in only 9 of these counties was defoliation in excess of 50 percent during either year. In the remaining counties defoliation averaged less than 15 percent. Evidence of sawfly work was seen in only one of the plots in Wisconsin (near Brule), and not at all in Michigan. Damage was observed on individual trees at Park Falls and Griffith State Nursery in Wisconsin, and at Merriweather in Michigan, but nowhere in these two States was the defoliation general. Infestation in Minnesota The State of Minnesota possesses almost two-thirds of the total area of tamarack type^ in the Lake States (Cunningham 1). Most of the area is in the northern tier of counties that lie within the present limits of sawfly infestation. Table 1 shows a breakdown by counties of the tamarack acreage within each defoliation zone (see fig. 1). The acreage of tamarack type for each county was supplied by the Lake States Forest Experiment Station. The sources of these estimates are the Forest Surveys for 1938 and surveys in progress. Figures for the Superior National Forest are based on the forest-management plan for 1936, and data for St. Louis and Lake Counties outside the National Forest boundaries on estimates provided by members of the Superior National Forest staff. 3/ Butcher, J. W. Studies on larch sawfly populations, factors con- tributing to their fluctuations and early recognition of host decadence. 1951. (Unpublished thesis University of Minnesota.) 4/ Stands containing 50 percent of swamp conifers with tamarack outweighing other species. -5- Table 1. --Estimated acreages of tamarack type infested with the larch sawfly in each defoliation zone (see fig. 1) in northern Minnesota County Zone 1 Zone 2 Zone 3 Zone 4 Total Lake of the Woods 62,000 _ , mm — 62,000 Koochiching 112,480 35,520 - - 148,000 Clearwater - - - 8,400 8,400 Beltrami 26,404 25,760 - 12,236 64,400 Hubbard - - 3,900 - 3,900 Wadena - - 1,500 - 1,500 Cass - 14,514 14,514 10,072 39,100 Crow Wing - - 8,528 1,872 10,400 Aitkin - - 16,154 23,246 39,400 Itasca - 23,932 - 14,668 38,600 Carlton - - 350 3,150 3,500 Lake and St. Louis 27,812 - - 11,919 39,731 Cook 89 - - - 89 Total 228,785 99,726 44,946 85,563 459,020 Control Experiments In most cases where successful control of forest-insect populations has been achieved through the use of chemicals, there have not been the complicating biological limitations to insecticide effectiveness that have been noted for the larch sawfly. One objective of the control experiments was to try to assess the importance of life-history traits that might enable some of the sawflies to escape contact with the insecticide unless more than one application was made. Quite apart from the separate effects of individual influences, however, it seemed desirable to determine the degree of control that could be achieved under prevailing conditions. Two main obstacles to high sawfly mortality from insecticides were apparent before the control experiments were started. One was the tendency for some of the overwintering population (prepupae) to remain in the soil for more than a year. The other was the relatively long period when sawfly adults were active, so that susceptible and nonsusceptible life stages overlapped, and at no time was the entire population vulner- able. The larval and adult stages were susceptible to sprays, but the egg and cocoon stages were not. In 1949 aerial spraying with technical DDT in fuel oil, at the rate of 1 pound per gallon per acre, was carried out on 80 acres of lightly in- fested tamarack on the Chippewa National Forest when most of the larvae were in the late instars. From the time of application until feeding -6 terminated, 81 percent more frass was produced in an adjacent check plot than in the treated plot. In 1950 it was decided to explore the possibility of securing better control with the same method by treating as near as possible to the peak of adult activity as well as of frass production by the larvae. For spraying the adults two areas known to be heavily infested in 1949 were selected- -one near Ebro, Minn., containing 90 acres, and the other near Bagley, Minn., containing 80 acres. The areas were treated on June 30, the date coinciding with the best available estimates of the actual peak of adult emergence. Each week throughout the season frass was collected in trays, 3 by 1 by 1/2 feet, set out in two sprayed and two check plots, one tray in each plot. The total collection for the season averaged 25.36+4.18 ml. from the sprayed plots and 38.62+4.44 ml. from the check plots. These figures indicate a 34.3-percent reduction in frass in sprayed plots on the assumption that both check and sprayed areas had the same populations before spraying. The differences between sprayed and unsprayed plots subsequent to spraying, while statistically significant, were so small that no recommendations for adult spraying seemed advisable. In 1950 a treatment to control the larvae was made in two blocks of heavily infested tamarack at Ely, Minn. Again the insecticide was DDT applied at the rate of 1 pound per gallon of fuel oil per acre. The appli- cation was made on August 2, approximately a week before the peak of larval feeding, although between 10 and 20 percent of the season's pop- ulation had cocooned by this time. The effect of the treatment was evaluated in the same manner as for the adult spraying, except that the trays were installed 3 days before treatment in order to check the uniformity of the sawfly population on the plots. The average volumes of frass per tray collected weekly for the check and sprayed plots are given below: Milliliters of frass Date Sprayed plots Check plots August 2 9 16 23 Up to the time of spraying, there was no statistically significant difference between frass volume from the check areas and that from the sprayed areas, which could be attributed to anything but sampling variation (Snedecor 4). One week later, however, there were signifi- cant differences, indicating two populations with unequal means, the differences due conceivably to spraying. Differences between the 6.8 9.3 20.5 82.1 3.8 35.3 0.1 5.0 - 7- seasonal frass accumulations were also significant and, assuming com- parable populations in check and sprayed plots prior to treatment, pop- ulation reduction based on frass-production figures was 77 percent in the sprayed plots. More work needs to be done to verify these results. However, the population reduction given above does seem consistent with the expected mortality, considering the obstacles to control that stem from the spread of life-history stages. This work has established the susceptibility of larch sawfly larvae to DDT and, within limits imposed by the circumstances under which the work was done, demonstrates the degree of control that could be expected under conditions prevailing in 1950. Conclusion In view of the results of these tests, the usefulness of DDT sprays to control larch sawfly outbreaks is open to question. If allowance is made for the 10 to 20 percent of the total population that cocooned, and for the fact that a reduction of only 77 percent of the larval population in the sprayed plots can be accounted for through reduced frass production, it is evident that mortality of the total resident population from a single spray application may not exceed 70 percent. The 30 percent surviving may be enough to perpetuate the outbreak. However, any conclusions on population reduction drawn from these experiments must take into con- sideration the high mortality that occurs from climatic and biological factors. The exact effect of these influences is not known, but it is believed that unfavorable moisture and temperature conditions cause mortality among overwintering prepupae. Sawfly survival is also in- fluenced somewhat by parasites, and to a greater extent by rodents. Although its permanent effect on the sawfly population may be of questionable value, spraying may be useful for minimizing damage to stands already weakened by repeated attacks. Apart from long-range sawfly control, protection from imminent defoliation could mean the difference between death or survival of a particular stand. In this connection knowledge of the susceptibility of larvae accessible to the insecticide is of value. The possible effect of a 70 -percent reduction in the larval population on the succeeding generations cannot be estimated, but observed differences between sprayed and unsprayed plots at Ely, Minn., the year after treatment indicate that such differences might be considerable under some conditions. Thus, even a partial reduction in sawfly populations might offer protection from severe defoliation for more than one season. UNIVERSITY OF FLORIDA - 8 - 3 1262 09239 6414 Literature Cited (1) Cunningham, R. N. 1950. Forest resources of the Lake States region. U. S. Forest Serv., Forest Res. Rpt. No. 1, 57 pp. (2) Hodson, A. C. 1940. Minnesota forest insect survey for 1940. Minn. Agr. Expt. Sta., Misc. Jour. Ser., Paper No. 448. (3) and Christensen, C. M. 1941-47. Minnesota forest insect and disease surveys. Minn. Agr. Expt. Sta., Sci. Jour. Ser., Paper Nos. 1997 (1941), 2084(1942), 2194(1943), 2296 ( 1945); Misc. Jour. Ser., Paper No. 634 (1947). (4) Snedecor, George W. 1946. Statistical methods. 485 pp. Ames, Iowa. (5) Swaine, J. M. 1933. The relation of insect activities to forest development as exemplified in the forests of eastern North America. Sci. Agr. 14(1): 8-31.